A series of numerical simulations are conducted (using GFDL CM2.1) to investigate the global mean precipitation and temperature change in response to climate variation. Experiments under different carbon dioxide (CO2) forcing indicate an obvious precipitation hysteresis. There is a significant linear relationship between global mean precipitation and surface temperature, but precipitation is also influenced directly by CO2 concentration. During the experiments in which CO2 concentration rises up and then falls back, precipitation change lags behind surface temperature, which leads to the precipitation hysteresis. While CO2 increasing, the enhanced greenhouse effect will lead to immediate intension of atmospheric long-wave absorption, which will bring net radiative energy income to atmosphere. To balance the energy budget, upward latent heat has to be restrained, so the additional CO2 has inhibiting effect on precipitation. The subsequent warming mainly induces increasing in outgoing long wave radiation at TOA and backward long wave radiation at surface, which is equivalent to a radiative cooling for atmosphere, and then causes precipitation rising. While CO2 decreasing, on the contrary, the subdued greenhouse effect tends to intensify precipitation and the temperature reduction will reduce the precipitation. Different effects on precipitation from temperature and CO2 determine the precipitation hysteresis.